Discover Aircall AI
What is Aircall AI?
Aircall AI combines telephony infrastructure and AI capabilities in one developer platform, helping you build AI-powered voice and messaging experiences into your products faster without managing speech-to-text, LLM pipelines, call routing, or telephony operations yourself.
With Aircall AI, you can extract structured insights from conversations, automate customer workflows, and launch AI-powered voice experiences on top of the calling platform your customers already use.
Today, you can build with two products
Structured outputs derived from every call: transcripts, summaries, sentiments, topics, action items, playbook results, and call evaluations.
A conversational agent that runs calls and messaging end-to-end, interacts with external systems, follows workflows, and uses knowledge bases.
You can learn more about how Aircall AI works in our product help center
AI Assist vs. AIVA
Aircall AI supports two patterns. Depending on your product and your users, one of them might be the right fit on its own — or both together might be. They're designed to coexist in the same integration.
| AI Assist | AIVA | |
|---|---|---|
| What it does | Turns every call and message into structured intelligence. | Runs calls and conversations on your behalf. Takes actions with external systems |
| When it runs | After (or during) human conversations | During the conversation itself |
| What you get | Transcripts, summaries, sentiment, action items, QA results | Conversation outcomes + structured answers from the call. Custom actions. |
| How you use it | Sync insights into your product via API/webhooks | Trigger conversations from your system and act on results. Provide tools |
| API | Conversation Intelligence API | AI Voice Agent API |
How it works in practice
AI Assist
For AI Assist, the model is a classic REST API pattern: you react to events and update your systems with structured outputs (transcripts, summaries, sentiment, action items). It fits naturally into event-driven workflows in your product.
AI Voice Agent (AIVA)
For AIVA, it’s a different paradigm. Instead of predefining orchestration and logic, you expose capabilities, and AIVA uses them dynamically during the conversation, guided by high-level instructions defined per agent by your customers.
Inbound AIVA calls
Everything happens inside Aircall. You configure AIVA in Aircall — define its context, instructions, and the actions it can use.
Your main responsibility as a developer/partner is making sure your system is ready to support those actions, either via an MCP connection (for approved partners) or a capable REST API. Once that’s in place, AIVA can handle the full conversation end-to-end.
Outbound AIVA calls
Control shifts to your product. You essentially orchestrate the conversation from your system, and AIVA executes it. You decide:
- When to trigger the call
- Which AIVA agent to use
- What context, flow, and parameters to pass
You can find a full step-by-step guide for outbound calls in the dedicated tutorial — this section is just to give you the high-level mental model within Aircall AI.
Updated 3 days ago
What’s Next
Now that you know a bit more about our AI offerings. See what you can do with them